More on digital cameras

Jan 8, 2007 15:50 GMT  ·  By

Today, we are going to take a closer look at how digital cameras work. This article will concentrate on color capturing techniques, resolution issues and a few important things about image exposer and focus.

Perfect Colors Remember the minute photosites on the CDDs? As a matter of fact, these minuscule devices are colorblind. They only keep track of the total intensity of the light that strikes their surface. In order to get a full color image, CCD or CMOS sensors use filtering techniques to split the light in its three primary colors. As soon as the camera records all three colors, a combination process of this information begins and the camera's hardware is able to reproduce wide color ranges.

The Bayer filter is to be found in many of today's compact digital cameras. This type of filter is used in combination with an optical anti-aliasing filter to reduce the aliasing due to the reduced sampling of the different primary-color images. Afterwards, a demosaicing algorithm is used to interpolate color information and thus create a full array of RGB image data. This is an economical and practical way to record the primary colors. The Bayer filter array is permanently placed over each individual photosite. It essentially breaks up the light from the sensor into a variety of red, blue and green pixels. Through this process emerges the possibility of obtaining enough information in the general vicinity of each sensor in order to make very accurate guesses about the true color at that location. This "looking" at the other pixels in the neighborhood of a sensor and making a maths-based guess is known as interpolation.

The Bayer filter pattern is a repeating 2?2 mosaic pattern of light filters, with green ones at opposite corners and red and blue in the other two positions. The higher amount of green (twice as the blue and the red ones combined) is in concordance with the properties of the human visual system, which determines brightness mostly from the green color, being far more sensitive to brightness than to hue or saturation. There also exist 4-color filter patterns, which involve two different hues of green. This array provides potentially more accurate color ranges, but requires additional interpolation processes.

Digital cameras use specialized demosaicing algorithms to convert Bayer pattern mosaic into an equally sized mosaic of true colors that can be recognized by the human eye. The key is that each colored pixel can be used more than once. The true color of a single pixel can be determined by averaging the values from the closest surrounding pixels via specific interpolation techniques.

Firmware in the camera ultimately interprets the raw data from the sensor to obtain a full color image. This is because the RGB color model requires three intensity values for each pixel: one for each red, green, and blue primary colors (other color models such as the CMYG requires 4 values per pixel for each of the cyan, magenta, yellow and green colors). A single sensor element cannot simultaneously record the various intensities, and so a color filter array (CFA) must be used to selectively filter a particular color for each pixel.

The advantages of this method are that only one sensor is required, and all the color information (red, green and blue) is recorded simultaneously. This translates into smaller, cheaper cameras that can be useful in a wider variety of situations. Some single-sensor cameras may incorporate an alternative to the Bayer filter pattern. The most important alternative is the X3 technology, which embeds red, green and blue photodetectors into silicon chips. And with this technology we jump to high-end cameras.

High-end and professional cameras usually use more than one CCD. In this case, one would encounter beam-splitter single-shot 3CCD approaches, three-filter multi-shot approaches, or Foveon X3 sensors that do not use anti-aliasing filters, nor demosaicing algorithms. The highest quality cameras use three separate sensors, each with a different filter. A beam splitter directs light to the different sensors dividing equal amounts of light for each CDD and thus, the three sensors get an identical look at the image. However, each CDD has a red, blue or green filter, and can only respond to one of these primary colors.

Another high-end method is to include a rotating array of red, blue and green filters in front of a single sensor. In this case, the sensor records three separate images in rapid succession. This solution is hindered by the annoying latencies as the three images aren't taken at precisely the same moment, requiring both the camera and the target of the photo to remain stationary for the full process.

The advantage of these high-end methods is that the camera records each of the three colors at each pixel location, making for more accurate color ranges. Nonetheless, cameras that use these technologies are bulkier and cost much more than compact ones.

Legions of pixels By now we know that each photosite on a CCD produces a pixel in the final image, and a demosaicing/interpolation algorithm is needed to turn the image with only one wavelength range per pixel into an RGB image where each pixel carries three values to represent a complete true color.

The one attribute most commonly compared on cameras is the pixel count or the total resolution. Due to the ever increasing sizes of sensors, the pixel count is expressed into millions. A million pixels is accordingly known as a megapixel. It is commonly presumed that the only important aspect of the camera is the total resolution. This is far from accurate as the quality of a digital image is also influenced by a series of other factors such as sensor size, lens quality, and the organization of the pixels (it is interesting to point out that a monochrome camera without a Bayer filter mosaic has a higher resolution than a typical color camera). Recent compact digital cameras that include high-resolution image sensors are oftentimes criticized for having so small CCDs that their resolution is greater than the lens could possibly deliver.

Moreover, excessive pixels can even lead to a decrease in image quality. And with ever-shrinking pixel sensors that capture fewer photons, we see the signal-to-noise ratios decreasing. This means that the images get quite noisy, featuring poor shadow region quality and generally poorer overall quality.

However, as the technology has improved, benefits for the consumers have also emerged. Take for instance the costs, which have decreased dramatically. Measuring the "pixels per dollar", there has been a continuous and steady increase when talking about the number of pixels each dollar buys in a new digital camera. This trend has been dubbed "Hendy's Law" as a specific incarnation of Moore's Law.

Focus, dude, focus! Digital cameras have to control the amount of light that reaches the sensor. In order to achieve good results, the digital camera borrows two important features from conventional cameras:

- Aperture - this is the size of the opening that allows light to get in the camera. The aperture is automatic and can't be changed in most digital cameras, but DSLRs allow manual adjustment to give professionals and hobbyists more control over the final image. - Shutter speed - is the amount of time during which light can pass through the aperture. Unlike film, the light sensor in a digital camera can be reset electronically, so digital cameras have a digital shutter rather than a mechanical shutter.

These two aspects work together to deliver what is known in photographic terms as the exposure of the sensor. In addition to controlling the amount of light, the camera features automated mechanisms so that the lenses are adjusted to control how the light is focused on the sensor. Generally, the lenses on digital cameras are very similar to conventional camera lenses and some digital cameras even use conventional lenses to decrease the costs. Compact digital cameras use automatic focusing techniques, while DSLRs feature interchangeable lenses that can be manually focused in addition to auto-focusing and programmable functions.

The focal length is one of the most important differences between the lens of a digital camera and the lens of a 35mm camera. The focal length represents the distance between the lens and the surface of the sensor. CCD and CMOS sensors from different manufacturers vary in size, but still they're smaller than a piece of 35mm film. In order to project the image onto a smaller sensor, the focal length has to be shortened by the same proportion.

Focal length is also important in determining the magnification factor, or zoom, when you look through the camera. For 35mm cameras, the usage of a 50mm lens provides a natural view of the subject. By increasing the focal length, we also increase the magnification, and objects appear to get closer. By reversing the focal length, we end up with a zoom out effect. A zoom lens has an adjustable focal length, and digital cameras can have optical or digital zoom features (some compact and prosumer-grade cameras have both features). Some cameras also have macro focusing capabilities, meaning that the camera can take pictures from a minimal distance from the subject. Digital cameras usually feature one of the following type of lenses:

a)Fixed-focus, fixed-zoom lenses - the kind of lenses that you would find on disposable and inexpensive cameras: great for snapshots, but fairly limited hen it comes to overall quality. b)Optical-zoom lenses with automatic focus - similar to the lenses on a traditional video camcorder, these have "wide" and "telephoto" options and automatic focus. . These actually change the focal length of the lens rather than just digitally magnifying the information that hits the sensor. c)Digital-zoom lenses - With digital zoom, the camera takes pixels from the center of the image sensor and interpolates them via the firmware to make a full-sized image. Depending on the resolution of the image and the sensor, this approach may create a grainy or fuzzy image, drastically reducing the quality of the picture. d)Replaceable lens systems - These are similar to the replaceable lenses on a 35mm camera. This is the domain of the DSLR professional cameras as they can easily use a wide variety of existing 35mm camera lenses.

Optical zoom lens with autofocus
Replaceable lens system

I think that would suffice in the case of our in depth look at digital cameras. Tomorrow we'll analyze some general aspects about digital video cameras.

Photo Gallery (2 Images)

Open gallery